diff options
author | Jeff Darcy <jdarcy@redhat.com> | 2015-01-06 10:03:49 -0500 |
---|---|---|
committer | Vijay Bellur <vbellur@redhat.com> | 2015-01-09 10:04:11 -0800 |
commit | 548547b2e41c8e2cf79b929405cf18aecbdedebc (patch) | |
tree | 8dba5d41c08edf366244e6679157419c999b1762 /xlators/protocol/auth | |
parent | 9408dc7b416ca80b3b8d8ecae2ef75c7e9cd21cd (diff) |
transport: fix default behavior for SSL authorization
Previously, enabling SSL authentication/encryption but not authorization
required explicitly setting ssl-allow=*. Now that same behavior is the
default (i.e. when ssl-allow is not set).
Also, there's no reason that a name used for *login* auth (typically a
UUID for internal purposes or a human name when using SSL) should
validate as an RFC-compliant host name or IP address. Therefore the
validation only occurs when the auth type is "addr" (not "login" or
anything else).
Change-Id: I01485ff4f0ab37de4b182858235a5fb0cf4c3c7d
BUG: 1179208
Signed-off-by: Jeff Darcy <jdarcy@redhat.com>
Reviewed-on: http://review.gluster.org/9397
Reviewed-by: Krishnan Parthasarathi <kparthas@redhat.com>
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Vijay Bellur <vbellur@redhat.com>
Diffstat (limited to 'xlators/protocol/auth')
-rw-r--r-- | xlators/protocol/auth/login/src/login.c | 23 |
1 files changed, 22 insertions, 1 deletions
diff --git a/xlators/protocol/auth/login/src/login.c b/xlators/protocol/auth/login/src/login.c index 56b93a9f9e9..b53c5ccba21 100644 --- a/xlators/protocol/auth/login/src/login.c +++ b/xlators/protocol/auth/login/src/login.c @@ -38,7 +38,6 @@ auth_result_t gf_auth (dict_t *input_params, dict_t *config_params) gf_log ("auth/login", GF_LOG_INFO, "connecting user name: %s", username_data->data); using_ssl = _gf_true; - result = AUTH_REJECT; } else { username_data = dict_get (input_params, "username"); @@ -80,6 +79,28 @@ auth_result_t gf_auth (dict_t *input_params, dict_t *config_params) if (allow_user) { gf_log ("auth/login", GF_LOG_INFO, "allowed user names: %s", allow_user->data); + /* + * There's a subtle difference between SSL and non-SSL behavior + * if we can't match anything in the "while" loop below. + * Intuitively, we should AUTH_REJECT if there's no match. + * However, existing code depends on allowing untrusted users + * to connect with *no credentials at all* by falling through + * the loop. They're still distinguished from trusted users + * who do provide a valid username and password (in fact that's + * pretty much the only thing we use non-SSL login auth for), + * but they are allowed to connect. It's wrong, but it's not + * worth changing elsewhere. Therefore, we do the sane thing + * only for SSL here. + * + * For SSL, if there's a list *you must be on it*. Note that + * if there's no list we don't care. In that case (and the + * ssl-allow=* case as well) authorization is effectively + * disabled, though authentication and encryption are still + * active. + */ + if (using_ssl) { + result = AUTH_REJECT; + } username_cpy = gf_strdup (allow_user->data); if (!username_cpy) goto out; |